agent-skills-installer

Safely refactor .NET / C# code at Senior Engineer level — diagnose code smells, classify risk (SAFE/RISKY/DANGEROUS), check the test safety net (or add characterization tests first), apply smallest-change-at-a-time for one smell, preserve behavior, match project convention. Use whenever the user wants to actually rewrite, restructure, clean up, or improve existing code — phrases like refactor this, refactor code, clean up, restructure, improve code quality, fix code smell, extract function, extract class, rename, inline, simplify, make this cleaner, make this DRY. Also trigger after a dotnet-code-review when the user says "apply the fixes". Skill DOES modify code (unlike dotnet-code-review which only inspects).

Multi-dimensional .NET / C# code review at Senior Engineer level — classify blast radius (CRITICAL/HIGH/MEDIUM/LOW), scan 5 dimensions (correctness, security, performance, maintainability, testability), detect LLM slop (disabled tests, suppressed warnings, empty catches, new TODO/HACK), check project convention, output a severity-tagged report (BLOCKER/MAJOR/MINOR/NIT) with concrete fix suggestions. Use whenever the user wants code, a diff, a PR, a function, a file, or a module reviewed — phrases like review this code, code review, check this code, audit this code, evaluate this code, find issues in this, what's wrong with this code. Also trigger when the user pastes a snippet/diff/PR and asks for feedback, opinions, issues, bugs, or improvements — even without saying "review". Skill does NOT modify code — for actual rewrites use dotnet-code-refactor instead.

Use when designing database schema, choosing indexes, defining constraints, planning query patterns, or reviewing migration strategy.

Use when user asks to refactor, clean up, simplify, or restructure code. Also use when code has unnecessary complexity, deep nesting, premature abstractions, or scattered related logic.

Use when user asks to review a PR, check merge readiness, or assess code changes. Also use when given a PR URL or diff to evaluate.

Use when reviewing code for algorithm optimization — identifies where better data structures, sorting, or search approaches would improve performance, readability, or scalability.

Use when designing or reviewing APIs — covers resources, contracts, error handling, versioning, and examples.

Use when designing system architecture, planning new services, evaluating tradeoffs, or mapping data flow and rollout plans.

Use when generating changelog entries from commits or merged changes — produces user-facing and developer-facing formats.

Use when reviewing code for bugs, maintainability, correctness, edge cases, and architecture issues.

Use when debugging failures — identifies symptoms, traces root causes, and proposes concrete fixes with validation steps.

Use when writing technical documentation — setup guides, ADRs, API references, and usage instructions.

Use when user reports a bug, pastes an error message, says code isn't working, asks why something fails, or requests debugging help. Use even for vague reports like "fix this", "not working", "why is this failing", or any stack trace. Ensures deep context analysis before proposing a fix instead of guessing.

Use when planning or reviewing data or schema migrations — covers safety, rollback strategy, backward compatibility, and observability.

Use when reviewing performance bottlenecks, profiling hotspots, evaluating measurement strategy, or assessing optimization tradeoffs.

Use when reviewing code for security risks — trust boundaries, secrets handling, input validation, and common abuse cases.

Use when designing or writing tests — covers critical flows, edge cases, and regression risks.

Use when building UI — applies design tokens, color palettes (light/dark), typography, spacing, UX principles, and SEO best practices automatically regardless of framework.

Use when writing, modifying, or generating any code — automatically applies senior-level engineering principles regardless of programming language.

Use when handling production incidents, reliability concerns, on-call triage, infrastructure issues, monitoring, alerting, and post-mortem analysis.